In a field study with 69 subjects, we investigated the influence of smoking, exposure to environmental tobacco smoke (ETS), diet, and location of residence on biomarkers for polycyclic aromatic hydrocarbons (PAH), including urinary excretion of 1-hydroxypyrene and benzo[a]pyrene (BaP) adducts of hemoglobin and albumin. The self-reported smoking status and the extent of ETS exposure were verified by urinary cotinine measurements. ETS exposure was quantified by nicotine and 3-ethenylpyridine measurements on personal samplers worn by the nonsmokers over 5 or 7 days before blood and urine samples were collected. Smokers (n = 27), on average, excreted 0.346 microg/24 h 1-hydroxypyrene, whereas the corresponding value for nonsmokers (n = 42) was 0.157 microg/24 h. Average BaP adduct levels with hemoglobin and albumin were 0.105 fmol/mg and 0.042 fmol/mg, respectively, for smokers, and 0.068 fmol/mg and 0.020 fmol/mg, respectively, for nonsmokers. The differences, except for the hemoglobin adducts, were statistically significant. Of the 42 nonsmokers, 19 were classified as passive smokers. There was no significant difference in the PAH biomarkers between nonsmokers exposed to ETS and those not or rarely exposed to ETS. Total dietary BaP intake, as calculated from questionnaire data, did not correlate with any of the PAH biomarkers (r < 0.1). Subjects living in the suburbs tended to have higher BaP-protein adduct levels than subjects living in the city. Our findings suggest that diet and smoking are major sources for PAH exposure of persons not occupationally exposed to PAH, whereas the influence of ETS exposure is negligible. The lack of correlation between the dietary PAH intake and the PAH biomarkers may be due to the inaccuracy of the estimate for the dietary PAH intake.